A Mechanistic Model of H2S Corrosion of Mild Steel

Paper No.
07655
A MECHANISTIC MODEL OF H2S CORROSION OF MILD STEEL
Wei Sun and Srdjan Nesic
Institute for Corrosion and Multiphase Technology
Department of Chemical Engineering
Ohio University
342 West State Street
Athens, Ohio 45701
ABSTRACT
Glass cell experiments were conducted to investigate the mechanism and kinetics of
mild steel corrosion in H2S environments which is accompanied by iron sulfide scale
formation. By using the weight change method, the rates of corrosion and scale formation
were found. It was also concluded that mackinawite is the predominant type of iron
sulfide that formed on the steel surface under the test conditions studied, probably by a
direct reaction of H2S with the underlying steel. Based on the experimental results, a
mechanistic model of uniform H2S corrosion of mild steel is presented that is able to
predict the rate of corrosion with time. In the model, the corrosion rate of mild steel in
H2S corrosion is a function of H2S gas concentration, temperature, velocity, and the
protectiveness of the mackinawite scale. The amount of scale retained on the steel surface
depends on the scale formation rate as well as the scale damage rate. The scale formation
may occur by corrosion and/or precipitation, while the scale damage can be by
mechanical or chemical means.
Keywords: mechanism, kinetics, mackinawite, mechanistic model, solid state reaction,
H2S corrosion
Copyright
Government work published by NACE International with permission of the author(s). The material presented and the views expressed in this paper are
solely those of the author(s) and are not necessarily endorsed by the Association. Printed in the U.S.A.
1
INTRODUCTION
Internal CO2 corrosion of mild steel in the presence of hydrogen sulfide (H2S)
represents a significant problem for the oil and gas industry1-6. Surface scale formation is
one of the important factors governing the corrosion rate. The scale growth depends
primarily on the kinetics of the scale formation. As a part of a larger project focusing on
the formation of both iron carbonate and iron sulfides in CO2/H2S corrosion, the kinetics
of iron carbonate has been quantified and reported in a recent publication7. In contrast to
relatively straightforward iron carbonate precipitation in pure CO2 corrosion, in an H2S
environment, many types of iron sulfides1-5, 8, 9 may form such as amorphous ferrous
sulfide, mackinawite, cubic ferrous sulfide, smythite, greigite, pyrrhotite, troilite, and
pyrite, among which mackinawite is considered to form first on the steel surface by a
direct surface reaction1-4. The unknown mechanisms of H2S corrosion makes it difficult
to quantify the kinetics of iron sulfide scale formation. Therefore, in this study, the
mechanism of H2S corrosion as well as iron sulfide formation is investigated and a model
of the overall process is proposed in this paper.
EXPERIMENTAL PROCEDURE
The experiments which served as a basis for the development of the model were
already described in the previous publication.10 Here, only the most important features of
the experimental program will be mentioned to facilitate following of the text below. The
experiments were performed at atmospheric pressure in glass cells filled with a 1 wt.%
NaCl aqueous solution continuously purged with a mixture N2 (>99.9 vol%) and H2S
gases (H2S concentrations in the gas inlet: 0.0075 vol% to 10 vol%). The H2S
concentrations in both the gas phase (in ppmw and Pa) and the liquid phase (in mol/l)
under different test conditions are shown in Table 2. Rectangular and cylindrical
specimens made from X65 pipeline steel were exposed for 1-24h at 25-80oC. The
chemical composition of the X65 steel used for all the experiments is shown in Table 1.
Weight change measurement was used to obtain both the corrosion rate of mild steel and
the retention rate of iron sulfide scale.
SUMMARY OF THE EXPERIMENTAL RESULTS
It was observed and previously reported 10 that in pure H2S corrosion of mild steel
there was no significant effect of dissolved Fe2+ concentration on neither the corrosion
rate nor the iron sulfide scale retention rate. This was in sharp contrast with pure CO2
corrosion where the iron carbonate scale formation rate is a strong function of dissolved
Fe2+ concentration i.e. it depends heavily on iron carbonate supersaturation, which is a
major driving force for iron carbonate scale formation by precipitation.7 Actually it was
long known that iron sulfide films form even in solutions which are well undersaturated,1
i.e. at pH much lower than pH5.0-5.5 which was used in this study. In addition, the
structure and morphology of the iron sulfide films formed in H2S corrosion (which was
2
identified primarily as mackinawite) is very different from the iron carbonate films
formed in CO2 corrosion. One observes layered crystalline iron sulfide films, with cracks
and delaminations, often with the imprint of the underlying metal surface clearly visible
even after long exposures.10 Therefore it is hypothesized here that iron sulfide films
observed in the experiments form primarily by a direct heterogeneous chemical reaction
between H2S and iron at the steel surface (often referred to as a “solid state reaction”).*
This hypothesis does not exclude the possibility of iron sulfide films forming by
precipitation in supersaturated solutions over long periods of time, however in the
relatively short duration experiments the main mechanism of iron sulfide formation is the
direct chemical reaction between H2S and the steel surface. Even more importantly it is
thought that the thin and tight iron sulfide films formed in this way are one of the most
important controlling factors in H2S corrosion.
Effect of H2S concentration
A number of experiments were conducted to investigate the effect of H2S gas
concentration on the mackinawite scale formation in the solutions with H2S/N2 at the
temperature of 80oC. Figure 1 shows the comparison of corrosion rate and scale retention
rates expressed in the same molar units vs. H2S gas concentration after a 1 hour exposure.
The value for the scaling tendency which is the ratio of the two rates is also shown. The
comparison indicates that both the corrosion rate and scale retention rate increase with
the increase of H2S gas concentration, however, the corrosion rate is always higher than
the scale retention rate. The scaling tendency under the test conditions indicates that
between 40% and 72% of the iron consumed by corrosion ended up as iron sulfide on the
steel surface, with the balance lost to the solution. As a very small increase in the
dissolved Fe2+ was measured in the solution it was concluded that some of the iron
sulfide that formed on the steel spalled off in a spontaneous process probably due to
intrinsic growth stresses (since no flow was present in these experiments that would
impose extrinsic hydrodynamic stresses). In Figure 2 the same kind of data is presented
for a 24 hour exposure where a broader range of H2S gas concentrations was used: 0.0075
vol% – 10 vol%. The same conclusions apply as for the 1 hour exposure with the
exception that the magnitude of both the corrosion rate and scale retention rate is almost
an order of magnitude lower after 24 hours. Interestingly, the scaling tendency remains in
approximately the same range 33-70% suggesting that between one and two thirds of the
iron sulfide that is formed by the corrosion process is lost to the solution by spalling.
The reduction in reaction rate with time is accentuated by the direct comparison of
the 1-hour and 24-hour corrosion rates (Figure 3) and scale retention rates (Figure 4) at
different H2S gas concentration. Clearly the iron sulfide scale that is retained on the
surface over time becomes gradually more protective.
Effect of temperature
The effect of temperature on both the corrosion rate and the scale retention rate is
shown in Figure 5 for a 1 hour exposure and in Figure 6 for a 24 hour exposure at 1 vol%
H2S gas concentration. Very weak temperature dependence is observed even for the
*
This hypothesis is not entirely new, it has been mentioned a number of times in various publications on
H2S corrosion of steel. 1-4
3
shorter term exposure which all but disappears for the longer exposure times. The same is
obtained in experiments at H2S gas concentrations of 10 vol%, as shown in Figure 7 and
Figure 8. This seems to suggest that the corrosion rate is predominantly controlled by the
presence of the iron sulfide scale, with the effect increasing over time.
Effect of flow rate
The effect of flow rate has been investigated by varying the rotation rate of the
cylindrical working electrode up to 8000 rpm which corresponds to a peripheral velocity
of approximately 4 m/s and a wall-shear stress of 57 Pa, in experiments done with 0.04
vol% of H2S in the gas phase. The corrosion rate as a function of reaction time at
different velocities is shown in Figure 9. The corrosion rate clearly increases with
velocity and the effect is much more pronounced for shorter exposure times. For longer
exposures in flowing conditions, the corrosion rates decrease significantly just as they did
in experiments conducted under stagnant conditions, due to a buildup of a protective iron
sulfide scale. However, as shown in Figure 10, the scaling tendency which is on average
50%, in stagnant conditions decreases to below 20% under flowing conditions. This
suggests that a much larger fraction of the iron sulfide scale formed in the corrosion
process is lost to the solution due to the hydrodynamic stresses induced by the flow. Iron
sulfide film dissolution could be excluded due to a slight supersaturation of the solution
with respect to mackinawite.
It has been reported 6, 11 that the presence of chloride ions may have an effect on the
H2S corrosion rate of the mild steel; however, this effect was not studied and will not be
discussed in the present paper.
MODEL OF H2S CORROSION
Physico-chemical model
There seems to be a consensus that mackinawite scale forms on the steel surface as
a product of H2S corrosion1-4, 8, 9. In this study mackinawite was also found to be the
dominant iron sulfide species, as previously described10. Clearly, other types of iron
sulfide film were observed in the past on steel surfaces attacked by H2S, particularly in
long exposures; however it is still unclear what effect the variation in film composition
may have on the corrosion rate.
Based on an analogy with iron carbonate formation in CO2 solutions and due to its
rather low solubility, mackinawite was also thought to form by a precipitation
mechanism10. While this is clearly a theoretical possibility, as argued above, mackinawite
formation via a direct heterogeneous chemical reaction with iron on the steel surface
seems to be the more relevant mechanism. Many pieces of evidence seem to support this
conclusion:
4
•
•
•
•
•
•
very high reactivity of H2S with iron, mackinawite scale has been shown to
form in extremely fast (less than seconds) 1, 2, 12, which is much faster than
what one would expect from typical kinetics of a precipitation process1;
formation of mackinawite scale in highly undersaturated solutions (pH2-3)
where it is thermodynamically unstable (soluble*) 1;
no effect of solution supersaturation level on the rate of mackinawite
formation10;
layered structure of mackinawite scale often containing cracks and
delaminations, with steel surface imprint visible even after rather long
exposures 10 (also see Figure 11);
amount of mackinawite scale always being smaller than the amount of iron
lost to corrosion of mild steel (expressed in molar units, see for example
Figure 1 - Figure 8) and a lack of substantial mackinawite scale formation on
stainless steel and other corrosion resistant alloys (see Figure 12), both
suggesting that the iron “source” in mackinawite is the metal itself, rather
than the bulk solution;
very similar structure and morphology of the mackinawite scale seen in high
temperature sulfidation of mild steel exposed to gaseous and hydrocarbon
environments13, 14, 15, where the precipitation mechanism is impossible.
If this is accepted as sufficient evidence, it can be concluded that the corrosion of
mild steel in H2S aqueous environments proceeds by a very fast direct heterogeneous
chemical reaction at the steel surface to form a solid adherent mackinawite scale. The
overall reaction scheme can be written as:
Fe(s ) + H 2 S ⇒ FeS (s ) + H 2
(1)
As both the initial and final state of Fe is solid, this reaction is often referred to as the
“solid state corrosion reaction”. The formed mackinawite scale may dissolve depending
on the solution saturation level. For the typical pH range seen in oilfield brines (pH 4-7)
the solution is almost always supersaturated with respect to iron sulfide and the
mackinawite scale does not dissolve, actually in long exposures it may grow slowly by
precipitation from the bulk16. If the pH is decreased below pH4 the dissolution rate will
increase to a point where in the range pH2-pH3 no mackinawite can be detected on the
steel surface.1
Even if aqueous H2S is a weak acid just like carbonic acid, the corrosion
mechanism proposed above differs in sequence from what is believed to happen to steel
*
A case can be made that the reasoning about solubility of iron sulfide based on conditions in the bulk is
invalid since at a steel surface, due to corrosion of iron, there always exists a somewhat higher pH and a
possibility to exceed the solubility limit, even in acidic solutions. In an extreme this would apply to any pH
(however low) as well as to other precipitating salts such as iron carbonate. In reality, iron carbonate films
are never observed at pH significantly below the solubility limit (based on bulk conditions) while iron
sulfide films are. This fact undermines the theoretically plausible argument about the importance of surface
conditions. In addition, basing arguments on a surface pH, which is virtually immeasurable, is not very
practical and is incompatible with the bulk of the chemical and electrochemical literature.
5
exposed to pure CO2 solutions in the same pH range (pH4-7). In CO2 corrosion of steel,
iron first dissolves to form aqueous Fe2+ which then may or may not precipitate at the
metal surface to form iron carbonate (e.g. below pH5 iron carbonate typically does not
form and above pH6 it is almost always there). In H2S solutions, steel corrosion proceeds
to first form a mackinawite scale which then may or may not dissolve.
This first layer of mackinawite that forms very fast is extremely thin (<< 1µm) and
is invisible to the naked eye and even by a typical SEM 12. However it is rather protective
and for example reduces a CO2 driven corrosion rate typically by an order of
magnitude12.
With increased exposure times, at high H2S concentrations and temperature, the thin
mackinawite film grows rapidly. It is still unclear whether this growth is supported by
H2S penetration through the crystalline layer (by solid state diffusion) or is it by ionic
conduction of S2- , HS- , Fe2+, etc. through the semiconductive mackinawite matrix.
Outward diffusion of ferrous species is consistent with an electrochemical iron
dissolution mechanism and a mackinawite continued growth at the outer film/solution
interface. The inward diffusion of sulfide species is consistent with the here proposed
direct chemical reaction mechanism (1) and leads to mackinawite formation at the inner
film interface with the steel. In both cases the mechanical integrity of the growing layer is
weakened. Outward migration of Fe2+ leaves “voids” at the metal/mackinawite interface
i.e. it “undermines” the film what manifests itself as poor “adhesion” of the film to the
steel. Inward diffusion of the sulfide species leads to internal stresses in the film as
described below.
In the latter scenario, the solid state corrosion reaction (1) keeps generating
mackinawite at the inner interface of the mackinawite film with the steel. This leads to
epitaxial stresses arising from the different crystalline structures of the source iron and
the iron sulfide that formed in its place13. What is more important, the solid FeS is
calculated to be 2.56 times more voluminous than the iron it replaced, at the
mackinawite/steel interface. This, so called Pilling-Bedworth ratio (PBR)13, leads to an
increase of internal compressive stresses in the mackinawite film. When the mechanical
limit of the mackinawite is exceeded micro-cracking of the film occurs, thereby relieving
the internal stresses and the process starts all over again. These micro-cracks, which most
likely occur at mackinawite grain boundaries, serve as preferred pathways for more rapid
penetration of sulfide species which fuel the solid state reaction (1) to go faster17. It is
expected that in some instances, at stress concentration points, large cracks in the film
may appear as shown in Figure 11. The sulfide species penetrate even more easily at
these locations to feed the corrosion reaction (1), which makes even more sulfide film at
those locations and causes even more internal stressing and film failure. It is not difficult
to see how this feed-forward scenario could lead to an exponential growth of the reaction
rate and localized corrosion. This scenario also offers an explanation for an apparently
odd occurrence in H2S corrosion: experimental observations indicate that pits are usually
full of iron sulfide and even have a cap of sulfide which is thicker than elsewhere on the
steel surface, as shown in Figure 13 provided by Brown18. This appearance is very
different from localized attack in CO2 corrosion where pits are bare while the
6
surrounding steel is covered with a protective film. Finally, in this scenario the hydrogen
gas evolved by the corrosion reaction (1) builds up at the steel/film interface as it diffuses
out through the mackinawite film with difficulty. This may lead to the retardation of the
atomic hydrogen recombination reaction and hydrogen penetration into the steel. Indeed,
the hydrogen built-up at the steel/film interface may even bubble out and cause further
damage to the mackinawite film. The last few points are purely hypothetical and were
discussed here only because they are consistent with proposed mechanism of H2S
corrosion of steel and the resulting iron sulfide film growth. As there is no direct
evidence for them in the short term experiments presented here, these hypotheses needs
to be directly confirmed in the future.
As the mackinawite film goes through the growth/micro-cracking cycle, it thickens.
As larger crack appear, whole layers of the film may partially delaminate from the steel
surface starting another cycle of rapid film growth underneath, as shown in Figure 14.
Over longer exposures, this cyclic growth/delamination process leads to a layered outer
sulfide scale which is very porous. As this outer scale grows it will spontaneously spall a
process assisted by flow. Notwithstanding, if the bulk solution is undersaturated
(typically at 3<pH<4) the outer porous mackinawite scale will dissolve away as fast as it
forms, what may happen even to tight inner mackinawite film at pH<3.1
In summary, in H2S corrosion of mild steel two types of mackinawite layers form
on the steel surface:
• a very thin (<<1µm) and tight inner film and
• a much thicker (1-10 µm) layered outer scale which is loose and very porous.
The outer scale may be intermixed with any iron sulfide or iron carbonate that may have
precipitated out given the right water chemistry and long exposure time, what would
change its properties and appearance. Both the inner mackinawite film and the outer scale
act as barriers for the diffusion of the sulfide species* fueling the solid state corrosion
reaction (1). This is in addition to the diffusion resistance through the aqueous mass
transfer boundary layer.
Mathematical model
Based on the experimental results and the description of the H2S corrosion process
presented above a mathematical model can be constructed. The key assumptions are:
• the corrosion process happens via a direct heterogeneous solid state reaction
(1) at the steel surface;
• there is always a very thin (<<1µm) but dense film of mackinawite at the
steel surface which acts as a solid state diffusion barrier for the sulfide
species involved in the corrosion reaction;
• this films continuously goes through a cyclic process of growth, cracking and
delamination, what generates the outer mackinawite scale;
*
In the authors’ opinion, the outward diffusion by the Fe2+ can be neglected as it is inconsistent with the
proposed solid state corrosion reaction (1) and would lead to a formation of a very different looking and
behaving sulfide film which is more akin to iron carbonate.
7
•
•
this outer scale grows in thickness (typically >1µm) over time and also
presents a diffusion barrier;
the outer scale is layered, very porous and rather loosely attached, over time it
peels and spalls, a process aggravated by the flow.
Due to the presence of the inner mackinawite film and possibly the outer scale it is
assumed that the corrosion rate of steel in H2S solutions is always under mass transfer
control. Based on the discussions above, a schematic of the H2S corrosion process is
shown in Figure 15.
One can write the flux of sulfide species due to:
•
convective diffusion through the mass transfer boundary layer
(
Flux H 2 S = k m , H 2 S cb , H 2 S − co , H 2 S
•
(2)
molecular diffusion through the liquid in the porous outer scale
FluxH 2 S =
•
)
DH 2 S ε Ψ
δ os
(c
o,H 2S
− ci , H 2 S
)
(3)
solid state diffusion through the inner mackinawite film
Flux H 2 S = AH 2 S e
−
BH 2 S
RTk
⎛ c i ,H 2 S
ln⎜
⎜ c s ,H S
2
⎝
⎞
⎟
⎟
⎠
(4)
where:
Flux H 2 S
is expressed in mol/(m2s),
k m ,H 2S
is the mass transfer coefficient for H2S in the hydrodynamic boundary layer,
k m ,H 2 S = 1.00 × 10 −4 in nearly stagnant condition, in m/s,
cb , H 2 S
is the bulk concentration of H2S in the liquid phase in mol/m3,
co , H 2 S
is the interfacial concentration of H2S at the outer scale/solution interface in
DH 2 S
ε
Ψ
ci , H 2 S
δ os
mol/m3,
is the diffusion coefficient for dissolved H2S in water, DH 2 S = 2.00 × 10−9 ,
in m2/s,
is the outer mackinawite scale porosity,
is the outer mackinawite scale tortuosity factor,
is the interfacial concentration of H2S at the inner scale/film interface in
mol/m3.
is the thickness of the mackinawite scale δ os = mos / ( ρ FeS A ) in m,
8
A
is the mass of the mackinawite scale in kg,
is the surface area of the steel in m2,
AH 2 S , B H 2 S
are the Arrhenius constants, AH 2 S = 1.30 × 10−4 mol/(m2s) and BH 2 S = 15500
Tk
J/mol,
is the temperature in Kelvin,
cs,H 2S
is the concentration of H2S on the steel surface and is set to 1.00 × 10 −7 in
mos
mol/m3.
In a steady state the three fluxes are equal to each other and are equal to the
corrosion rate CR H 2 S . By eliminating the unknown interfacial concentrations co , H 2 S and
ci , H 2 S from equations (2) to (4), the following equation is obtained for the corrosion rate
of steel due to H2S:
CR H 2 S = AH 2 S e
−
BH 2 S
RTk
⎛ δ 0 .5
1
+
c b ,H 2 S − CR H 2 S ⎜
⎜ D H S ε Ψ k m ,H S
2
2
⎝
ln
c s ,H 2 S
⎞
⎟
⎟
⎠
(5)
This is a nonlinear equation with respect to CR H 2 S which does not have a explicit solution
but can be solved by using a simple numerical algorithm such the interval halving method
or similar. These are available as ready-made routines in spreadsheet applications or in
any common computer programming language. The prediction for CRH 2 S depends on a
number of constants used in the model which can be either found in handbooks (such
as DH 2 S ), calculated from the established theory (e.g. k m , H 2 S ) or are determined from the
experiments (e.g. AH 2 S , B H 2 S , c s , H 2 S ). The unknown properties of the outer sulfide scale
change with time and need to be calculated as described below.
It is assumed that the amount of scale retained on the metal surface at any point in
time depends on the balance of:
•
•
scale formation (generated by spalling of the thin mackinawite film underneath it
and by precipitation from the solution), and
scale damage (by hydrodynamic stresses and/or by chemical dissolution)
SRR
SFR − {
SDR
{= {
scale
retention
rate
scale
formation
rate
(6)
scale
damage
rate
where all the terms are expressed in mol/(m2s). As in this study it was found that
precipitation of iron sulfide did not play a significant role, neither did chemical
dissolution of the scale it can be written:
9
SRR
SDRm
{ = CR
{ − {
scale
corrosion
retention
rate
rate
(7)
mechanical
scale damage
rate
Experiments have shown that even in stagnant conditions about half of the sulfide scale
that formed was lost from the steel surface by spalling, i.e. SDRm ≈ 0.5 CR . Furthermore,
the rate of scale removal in flowing conditions increased with velocity and one can write:
(
)
SDRm = 0.5 1+ c va CR
(8)
where c ( ≈ 0.55 ) and a ( ≈ 0.2 ) are experimentally determined constants for a rotating
cylinder flow geometry. Clearly more experimentation is required to determine how and
if they apply in pipe flow.
Once the scale retention rate SRR is known, the change in mass of the outer scale
can be easily calculated as:
∆mos = SRR M FeS A ∆t
(9)
where M FeS is the molar mass of iron sulfide in kg/mol, ∆t is the time interval in
seconds. The porosity of the outer mackinawite scale was determined to be very high
( ε ≈ 0.9 ), however due to its layered structure the tortuosity factor was found to be very
low Ψ = 0.003 .
A time-marching solution procedure could now be established where:
1. the corrosion rate CRH 2 S in the absence of sulfide scale can be calculated by
using equation (5), and assuming δ os = 0 ,
2. the amount of sulfide scale ∆mos formed over a time interval ∆t is calculated by
using equation (9),
3. the new corrosion rate CRH 2 S in the presence of sulfide scale can be calculated
by using equation (5),
4. a new time interval ∆t is set and steps 2 and 3 repeated.
A small complication arises from the fact that at very low H2S gas concentrations
(ppmw range) iron sulfide still forms and controls the corrosion rate; however the
corrosion process is largely driven by the reduction of protons.* In an analogy with the
approach laid out above, the convective diffusion flux of protons through the mass
transfer boundary layer is:
*
Similar is true in combined CO2/H2S corrosion of steel which is driven by CO2 but largely controlled by
the presence of iron sulfide films. Mixed CO2/H2S corrosion is not considered in this paper.
10
(
FluxH + = k m , H + cb , H + − co , H +
)
(10)
which in a steady state is equal to the diffusion flux through the pores of the iron sulfide
scale:
FluxH + =
DH + ε Ψ
δ oc
(c
o,H +
− ci , H +
)
(11)
which is equal to the solid state diffusion flux through the thin mackinawite film:
Flux H + = AH + e
−
B
H+
RTk
⎛c +
ln⎜ i ,H
⎜c +
⎝ s ,H
⎞
⎟
⎟
⎠
(12)
which is equal to the corrosion rate by protons CRH + . By eliminating the unknown
interfacial concentrations co , H + and ci , H + from equations (10) to (12), the following
expression is obtained for the corrosion rate driven by protons and controlled by the
presence of the iron sulfide scale:
CR H + = AH + e
−
B
H+
RTk
⎛ δ 0 .5
1
+
c b ,H + − CR H + ⎜
⎜ D +ε Ψ k +
m ,H
⎝ H
ln
c s ,H +
⎞
⎟
⎟
⎠
(13)
where
FluxH +
is expressed in mol/(m2s),
km, H +
is the mass transfer coefficient for protons in the hydrodynamic boundary
layer, k m , H + = 3.00 × 10 −4 in nearly stagnant condition, in m/s,
cb, H +
is the bulk concentration of H+ in the liquid phase in mol/m3,
co,H +
is the interfacial concentration of H+ at the outer scale/solution interface in
DH +
mol/m3,
is the diffusion coefficient for dissolved H+ in water, DH + = 2.80 × 10 −8 , in
ci , H +
AH + , B H +
m2/s,
is the interfacial concentration of H+ at the inner scale/film interface in
mol/m3.
are the Arrhenius constants, AH + = 3.9 × 10−4 mol/(m2s) and BH + = 15500
J/mol.
11
cs,H +
is the concentration of H+ on the steel surface and is set to 1.00 × 10 −7 in
mol/m3.
The total rate of corrosion is equal to the sum of the corrosion caused by H2S and
the corrosion caused by H+.
CR = CRH 2 S + CRH +
(14)
This completes the description of a basic mechanistic model of pure H2S corrosion
of mild steel. A number of important effects are not covered by the model as presented
above:
• corrosion due to CO2, organic acids, etc.;
• formation of scale due to precipitation of iron sulfide, iron carbonate, etc.;
• dissolution of the outer scale and inner film due to a very low pH;
• localized attack.
The first three effects have been already accounted for in a more comprehensive
CO2/H2S model (MULTICORP V4.0); however the presentation of this model exceeds
the scope of the present paper.
VERIFICATION AND TESTING OF THE MODEL
The model predictions are compared with the experimental results at different test
conditions. Figure 16 shows the comparison of the corrosion rate vs. the reaction time for
a series of experiments done at 80oC. One should keep in mind that the experimental
results are time-averages over 1 h and 1-24 h periods while the predictions represent
“instantaneous” corrosion rates. Clearly the model successfully captures the downward
trend of the corrosion rate with time as well as the undesirable effect of high H2S
concentrations on the general corrosion rate. Figure 17 shows the comparison of the
measured and predicted scale retention at different reaction times. The predicted scale
growth is rapid in the first few hours and then gradually levels off, leading to what is
often referred to as a “parabolic film growth regime”. After all the cases available in this
experimental study were simulated with the model, the comparison of the predicted H2S
corrosion rates and the measured values is shown in Figure 18. Overall one can claim
reasonable agreement keeping in mind the scatter in the experimental results.
The model was tested by making simulations outside the range of parameters used
in the experimental study described above, i.e. the model was used to extrapolate the
corrosion rates to higher partial pressures of H2S as well as much longer exposure times
(both are very complicated and expensive to achieve in a laboratory setting). In Figure 19
one can see the predictions ranging from partial pressure of H2S as low as 0.16 Pa in the
gas phase (what corresponds to 1.6 ppmw at 1 bar total pressure) all the way up to 2.7 bar
H2S partial pressure. The simulations were extended to 10 years and shown on a log-scale.
Clearly, the corrosion rate decreases to a very low value in all cases, while at the lowest
12
H2S concentration it may take less than a day at the highest it may take as long as few
years. The film thickness prediction is shown in Figure 20, indicating a scale thickness
which is only a few mm thick even at the highest H2S partial pressures and in very long
exposures.
CONCLUSIONS
The primary findings of this project are:
•
•
•
•
Mackinawite is the predominant iron sulfide formed on the steel surface, most
likely by solid state reaction.
The corrosion rate of mild steel in H2S corrosion is affected by H2S gas
concentration, temperature, velocity, and the protectiveness of the scale.
The scale retained on the steel surface depends on both the scale formation rate
and the scale damage rate. The scale formation rate includes both the corrosion
rate and precipitation rate. The scale damage rate includes the damages by both
mechanical removal and chemical removal.
A mechanistic model of H2S corrosion is developed to accurately predict the
H2S corrosion process.
ACKNOWLEDGEMENTS
During this work, W. Sun was supported by Ohio University Donald Clippinger
Fellowship. A part of the experiments in this paper were conducted in CANMET Materials
Technology Laboratory, Natural Resources Canada and the authors acknowledge Dr.
Sankara Papavinasam for his support. The authors would also like to acknowledge the
companies who provided the financial support and technical guidance for this project.
They are BP, Champion Technologies, Chevron, Clariant, ConocoPhillips, ENI,
ExxonMobil, MI SWACO, Nalco, Saudi Aramco, Shell, Total, Tenaris, Baker Petrolite,
Columbia Gas Transmission, and PPT.
13
REFERENCES
[1]
D. W. Shoesmith, P. Taylor, M. G. Bailey, & D. G. Owen, J. Electrochem. Soc., The
formation of ferrous monosulfide polymorphs during the corrosion of iron by
aqueous hydrogen sulfide at 21 oC, 125 (1980) 1007-1015.
[2]
D. W. Shoesmith, Formation, transformation and dissolution of phases formed on
surfaces, Lash Miller Award Address, Electrochemical Society Meeting, Ottawa,
Nov. 27, 1981.
[3]
S. N. Smith and E. J. Wright, Prediction of minimum H2S levels required for slightly
sour corrosion, Corrosion/94, Paper no. 11, NACE International, Houston, Texas,
1994.
[4]
S. N. Smith and E. J. Wright, Prediction of corrosion in slightly sour environments,
Corrosion/2002, paper no. 02241, NACE International, Houston, Texas, 2002.
[5]
S. N. Smith and M. Joosten, Corrosion of carbon steel by H2S in CO2 containing
oilfield environments, Corrosion/2006, Paper no. 06115, NACE International,
Houston, Texas, 2006.
[6]
M. Bonis, M. Girgis, K. Goerz, and R. MacDonald, Weight loss corrosion with H2S:
using past operations for designing future facilities, Corrosion/2006, paper no.
06122, NACE International, Houston, Texas, 2006.
[7]
W. Sun and S. Nesic, Basics revisited: kinetics of iron carbonate scale precipitation
in CO2 corrosion, Corrosion/2006, paper no. 06365, NACE International, Houston,
Texas, 2006.
[8]
P. Taylor, The stereochemistry of iron sulfides—a structural rationale for the
crystallization of some metastable phases from aqueous solution, American
Mineralogist. 65 (1980) 1026-1030.
[9]
J. S. Smith and J. D. A. Miller, Nature of sulfides and their corrosive effect on
ferrous metals: a review, British Corrosion Journal, 10 (1975) 136-143.
[10] W. Sun, S. Nesic and S. Papavinasam, Kinetics of iron sulfide and mixed iron
sulfide/carbonate scale precipitation in CO2/H2S corrosion, Corrosion/2006, paper
no. 06644, NACE International, Houston, Texas, 2006.
[11] B. F. M. Pots and S. D. Kapusta, Prediction of corrosion rates of the main corrosion
mechanisms in upstream applications, Corrosion/2005, paper no. 05550, NACE
International, Houston, Texas, 2005.
14
[12] K-L J. Lee and S. Nesic, EIS investigation on the electrochemistry of CO2/H2S
corrosion, paper no. 04728, NACE International, Houston, Texas, 2004.
[13] M. Schulte and Michael Schutz, The role of scale stresses in the sulfidation of steels,
Oxidation of Metals, 51 (1999) 55-77.
[14] A. Dravnieks, C. H. Samans, Kinetics of reaction of steel with hydrogen
sulfide/hydrogen mixtures, J. Electrochemical Society, 105 (1958) 183-191.
[15] D. N. Tsipas, H. Noguera, and J. Rus, Corrosion behavior of boronized low carbon
steel, Materials Chemistry and Physics, 18 (1987) 295-303.
[16] B. Brown, Srdjan Nesic and Shilpha Reddy Parakala, CO2 corrosion in the presence
of trace amounts of H2S, paper no. 04736, NACE International, Houston, Texas,
2004.
[17] M. Schutze, Protective Oxide Scales and Their Breakdown, John Wiley & Sons,
Chichester, 1997.
[18] B. Brown, H2S/CO2 corrosion in multiphase flow, Ohio University Advisory Board
Meeting Report, Internal report to be published, 2006.
15
TABLE
Table 1. Chemical Composition of X65 (wt.%) (Fe is the balance)
Al
As
B
C
Ca
Co
Cr
Cu
Mn
Mo
Nb
0.0032 0.005 0.0003 0.050 0.004 0.006 0.042 0.019 1.32 0.031 0.046
Ni
P
Pb
S
Sb
Si
Sn
Ta
Ti
V
Zr
0.039 0.013 0.020 0.002 0.011 0.31 0.001 0.007 0.002 0.055 0.003
Table 2. The concentrations of sulfide species at different concentrations of H2S
in the gas inlet stream and the H2S solution at pH 5 and Ptot 1 bar.
Temperature
/ oC
H2S concentration
in the gas / vol%
25
0.01
0.1
1
10
0.1
1
10
0.0075
0.015
0.024
0.04
0.1
1
10
60
80
16
c H 2 S( g )
ppmw
119
1189
11871
116512
1049
10471
102998
58
116
186
310
776
7752
76573
c H 2 S( g )
Pa
9.8
98
981
9808
81
813
8132
4
8
13
22
54
539
5392
c H 2 S( aq )
mol/l
9.4E-6
9.4E-5
9.4E-4
9.4E-3
4E-5
4E-4
4E-3
1.58E-6
3.16E-6
5.06E-6
8.44E-6
2.11E-5
2.11E-4
2.11E-3
FIGURES
0.1
Reaction rate / mol/h/m
2
Corrosion rate
Scale retention rate
% ST
67.8%
72.1%
40.9%
0.01
54.5%
10
%
1%
0.
1%
%
0.
04
4%
0.
02
5%
0.
01
0.
00
7%
0.001
H2S gas concentration
Figure 1. The comparison of corrosion rate (CR) and scale retention rate (SRR) in the
same molar units as a function of H2S gas concentration; ST=SRR/CR stands for Scaling
Tendency; total pressure p=1 bar, T=80oC, initial Fe2+ aqueous concentration: 0 ppm, pH
5.0-5.5, reaction time 1 hr.
0.01
2
Reaction rate / mol/h/m
57%
Corrosion rate
Scale retention rate
% ST
54.5%
45.7%
63.2%
70.4%
48%
0.001
33.8%
%
10
1%
0.
1%
%
0.
04
4%
0.
02
5%
0.
01
0.
00
7%
0.0001
H2S gas concentration
Figure 2. The comparison of corrosion rate (CR) and scale retention rate (SRR) in the
same molar units as a function of H2S gas concentration; ST=SRR/CR stands for Scaling
Tendency; total pressure p=1 bar, T=80oC, initial Fe2+ aqueous concentration: 0 ppm, pH
5.0-5.5, reaction time: 24 hr.
17
Corrosion rate / mm/yr
10
1hr
24hrs
1
0.1
0.04%
0.1%
1%
10%
H2S gas concentration
Figure 3. The corrosion rate vs. H2S gas concentration after 1 hr and 24 hr exposure at
total pressure p=1 bar, T=80oC, initial Fe2+ aqueous concentration: 0 ppm, pH 5.0-5.5.
Scale retention rate / g/h/m
2
10
1hr
24hrs
1
0.1
0.04%
0.1%
1%
10%
H2S gas concentration
Figure 4. The scale retention rate vs. H2S gas concentration after 1 hr and 24 hr exposure
at total pressure p=1 bar, T=80oC, initial Fe2+ aqueous concentration: 0 ppm, pH 5.0-5.5.
18
0.1
Reaction Rate / mol/h/m
2
CR H2S 1%
SRR H2S 1%
ST=67.8%
ST=56.8%
ST=78.3%
0.01
25
60
80
o
Temperature / C
Figure 5. The corrosion rate (CR) and scale retention rate (SRR) vs. temperature,
ST=SRR/CR stands for Scaling Tendency; conditions: total pressure p=1 bar, H2S gas
concentration is 1%, initial Fe2+ aqueous concentration: 0 ppm, pH 5.0-5.5, reaction time:
1 hr.
0.1
Reaction Rate / mol/h/m
2
CR H2S 1%
SRR H2S 1%
ST=23.3%
0.01
ST=51.7%
ST=70.4%
0.001
25
60
80
o
Temperature / C
Figure 6. The corrosion rate (CR) and scale retention rate (SRR) vs. temperature,
ST=SRR/CR stands for Scaling Tendency; conditions: total pressure p=1 bar, H2S gas
concentration: 1%, initial Fe2+ aqueous concentration: 0 ppm, pH 5.0-5.5, reaction time:
24 hr.
19
0.1
Reaction Rate / mol/h/m
2
CR H2S 10%
SRR H2S 10%
ST=72.1%
ST=60.9%
ST=63.6%
0.01
25
60
80
o
Temperature / C
Figure 7. The corrosion rate (CR) and scale retention rate (SRR) vs. temperature,
ST=SRR/CR stands for Scaling Tendency; conditions: total pressure p=1 bar, H2S gas
concentration: 10%, initial Fe2+ aqueous concentration: 0 ppm, pH 5.0-5.5, and reaction
time 1 hr.
0.1
Reaction Rate / mol/h/m
2
CR H2S 10%
SRR H2S 10%
ST=24.3%
ST=57.3%
0.01
ST=18.2%
0.001
25
60
80
o
Temperature / C
Figure 8. The corrosion rate (CR) and scale retention rate (SRR) vs. temperature,
ST=SRR/CR stands for Scaling Tendency; conditions: total pressure p=1 bar, H2S gas
concentration: 10%, initial Fe2+ aqueous concentration: 0 ppm, pH 5.0-5.5, and reaction
time: 24 hr.
20
6
Corrosion rate / mm/yr
0rpm
5
4000rpm
8000rpm
4
3
2
1
0
1
20
Time / hour
Figure 9. The corrosion rate vs. time for different rotational speeds; conditions: total
pressure p=1 bar, T=25oC, H2S gas concentration: 0.04%, initial Fe2+ aqueous
concentration: 0 ppm, pH 5.0-5.5.
Scaling tendency
100%
90%
velocity = 0rpm
80%
velocity = 4000rpm
velocity = 8000rpm
70%
60%
50%
40%
30%
20%
10%
0%
0
5
10
15
20
25
Reaction time / hour
Figure 10. The comparison of scaling tendency vs. reaction time under the conditions of
total pressure p=1 bar, T=25oC, H2S gas concentration 0.04%, initial Fe2+ aqueous
concentration 0 ppm, and velocity 0, 4000, and 8000rpm.
21
(a)
(b)
Figure 11. The film morphology showing polishing marks on the X65 mild steel (a)
1000X and (b) 5000X, under the conditions of total pressure p=1 bar, initial Fe2+ aqueous
concentration 0 ppm, H2S gas concentration 10%, T 60oC, reaction time 1 hour, pH
5.0-5.5, and velocity 0rpm.
(a-1)
(a-2)
(b-1)
(b-2)
Figure 12. The film morphology on the different steel surface (a-1) X65 mild steel Fe2+ 0
ppm, (a-2) X65 mild steel Fe2+ 50 ppm, (b-1) 316 stainless steel Fe2+ 0 ppm, (b-2) 316
stainless steel Fe2+ 50 ppm, under the conditions of total pressure p=1 bar, , H2S gas
concentration 0.1%, T 80oC, reaction time 24 hours, pH 5.0-5.5, and velocity 0rpm.
22
(a)
(b)
Figure 13. The morphology (a) and cross section (b) of the localize attack on the X65
mild steel surface in CO2/H2S environment under the conditions of Ptot 8bar, PH2S 8mbar,
PCO2 7.5bar, T 60oC, and the total reaction time is 10 days18.
Figure 14. Cross section of the scale formed on the X65 mild steel surface (at 1000X)
under the conditions of total pressure p=1 bar, initial Fe2+ aqueous concentration 0 ppm,
H2S gas concentration 1% (H2S/N2 gas), T 80oC, pH 5, the total reaction time is 24 hours.
23
Convective diffusion
Co
Cb
Molecular
diffusion
H2S
+
H
Ci
Solid state
diffusion
steel
water
Cs
Figure 15. A schematic of the H2S corrosion process.
Corrosion rate / mm/y
10
Exp. - 0.1% H2S
Model - 0.1% H2S
Exp. - 1% H2S
Model - 1% H2S
Exp. - 10% H2S
Model - 10% H2S
1
0.1
0.01
0
5
10
15
20
25
Time / hour
Figure 16. The experimental and prediction corrosion rate vs. time under the conditions
of total pressure p=1 bar, H2S gas concentration from 0.1% to 10%, T 80oC, reaction time
of 1 hour and 24 hours, pH 5.0-5.5, and velocity 0rpm.
24
1
Model - 0.1% H2S
Model - 1% H2S
Model - 10% H2S
0.1
SC / mol/m
2
Exp. - 0.1% H2S
Exp. - 1% H2S
Exp. - 10% H2S
0.01
0.001
0
5
10
15
20
25
Time / hour
Figure 17. The experimental results and predictions of the scale retention vs. time under
the conditions of total pressure p=1 bar, H2S gas concentration from 0.1% to 10%, T
80oC, reaction time of 1 hour and 24 hours, pH 5.0-5.5, and velocity 0rpm.
calculated corrosion rate / mm/yr
5
4
3
2
1
0
0
1
2
3
4
5
experimental corrosion rate / mm/yr
Figure 18. The comparison of the experimental corrosion rate and the calculated
corrosion rate under the conditions of total pressure p=1 bar, H2S gas concentration from
0.0075% to 10%, T 25oC, 60oC, and 80oC, reaction time of 1 hour and 24 hours, pH
5.0-5.5, and velocity from 0rpm to 8000rpm.
25
Corrosion rate / mm/y
1.E+01
1.E+00
1.E-01
H2S partial pressure
2.7 bar
1.0 bar
424 milibar
53 milibar
5.3 milibar
8 Pa
0.16 Pa
1.E-02
1.E-03
0.1
1
10
100
1000
Time / hour
1 day
1 month
10000
1 year
100000
10 years
Figure 19. Simulated corrosion rate as a function of time for a range of H2S partial
pressures; conditions T 80oC, pH 5, and stagnant.
H2S partial pressure
1
The thickness of the scale / mm
2.7 bar
1.0 bar
424 milibar
0.1
53 milibar
5.3 milibar
8 Pa
0.16 Pa
0.01
0.001
0.0001
0.1
1
10
100
1000
Time / hour
1 day
1 month
10000
1 year
100000
10 years
Figure 20. Simulated sulfide scale thickness as a function of time for a range of H2S
partial pressures; conditions: T 80oC, pH 5, and stagnant.
26